Searching and generating test inputs for mutation testing

Mutation testing is usually regarded as an important method towards fault revealing. Despite this advantage, it has proved to be impractical for industrial use because of its expenses. To this extend, automated techniques are needed in order to apply and reduce the method's demands. Whilst there is much evidence that automated test data generation techniques can effectively automate the testing process, there has been little work on applying them in the context of mutation testing. In this paper, search-based testing is used in order to effectively generate test inputs capable of revealing mutants. To this end, a dynamic execution scheme capable of introducing and guiding the search towards the sought mutants is proposed. Experimentation with the proposed approach reveals its superiority from the previously proposed methods. Additionally, the framework's feasibility and practicality of producing mutation based test cases are also demonstrated.